Model weights used by lipsync.py
================================

yunet.onnx  (face_detection_yunet_2023mar.onnx)
  Source : https://github.com/opencv/opencv_zoo/tree/main/models/face_detection_yunet
  Licence: MIT License, Copyright (c) 2020 Shiqi Yu. Commercial use permitted.

face_landmark.tflite / face_landmark.onnx  (MediaPipe Face Mesh v1)
  Source : https://storage.googleapis.com/mediapipe-assets/face_landmark.tflite
           (https://github.com/google-ai-edge/mediapipe)
  Licence: Apache License 2.0, Copyright Google LLC. Commercial use permitted.
  Model card: https://storage.googleapis.com/mediapipe-assets/Model%20Card%20MediaPipe%20Face%20Mesh%20V2.pdf
  face_landmark.onnx is a format conversion of the TFLite file made locally by
  setup_models.py (weights unchanged; fp16 weights widened to fp32).
